Hand Grip Strength in Relation to Morphological Measures of Masculinity, Fluctuating Asymmetry and Sexual Behaviour in Males And Females

Sex Hormones not only regulate reproductive function, but they also play a prominent role in the biology and physiology of several organs/tissues and in the pathophysiology of several diseases. During the last two decades, the information on the mechanisms of action of sex hormones, such as estrogens and androgens, has rapidly evolved from the conventional nuclear receptor dependent mechanisms to include additional non-nuclear, non-genomic and receptor-independent mechanisms. This highlights the need to update the current knowledge on sex hormones and their mode of action. Increasing evidence that exogenous/epigenetic factors can influence sex hormone production and action highlights the need to update our knowledge on the mechanisms involved. This book provides a systematic and updated overview of the male/female sex-hormones and their impact in the biology and physiology of various organs. Additionally, the book discusses their positive and negative association with the pathophysiology of various diseases (e.g. osteoporosis, cardiovascular-disease, hypogonadism, reproduction, cancer) and their therapeutic potential.

In this study we present data on relationships between HGS and i) objective measures of facial masculinity/femininity; ii) fluctuating asymmetry; iii) attractiveness and vi) sexual behaviour in a population of young males and females. The dataset presented here was not very large (total sample size of 100). Therefore, it cannot provide strong conclusions. Although the study of associations with HGS in this context is relatively new, quite some estimates have been published. It is therefore timely to review these results and combine them with the newly presented data here to come to more robust conclusions and suggestions for further research. We therefore also present a meta-analysis of all available data.

Study design
We measured bilateral asymmetry and masculinity/femininity from scans (HP scanjet G4050, 4800*9600 DPI) of hands and photographs (Nikon D70, 6 megapixel) of faces of 52 men and 48 women with an average age of 22.6 (SD = 2.66) and 22.3 (SD = 1.87) years respectively. The degree of handedness was also self-evaluated on a scale of 0 (extreme lefthanded) to 10 (extreme right-handed). Hand-grip strength (HGS) was determined using a Biometrics precision dynamometer. For each participant, the strength was determined twice on each side and the maximum value was obtained as HGS. All participants also completed a questionnaire asking for their age of first sexual contact and their total lifetime number of sexual partners. Each photograph was rated for its attractiveness by 10 to 30 opposite sex raters. As the repeatability of these ratings was about 30%, reliable estimates of attractiveness were obtained. For each participant the length of the left and right 2 nd (D2), 3 rd (D3), 4 th (D4) and 5 th digit (D5) as well as the width of the palm of the hand (P) were independently measured 3 times and averaged (Fig.1). On each photograph, initially 7 landmarks were placed on each side of the face to obtain measures of facial asymmetry: i.e.; the width of the eye (EW), the distance between the pupil of the eye and the widest point at the side of the nostrils (EN), and the distance between the cheek bone and the corner of the mouth (CM) (Fig.1). Landmarks were placed in 3 independent sessions (i.e., on three separate days) and distances were averaged across sessions to reduce measurement error. In addition, since traits within hands and face showed correlations in the signed FA, traits were averaged within hands and faces to obtain two composite estimates (handFA and faceFA1, see Van Dongen et al., 2009 for details). The relative lengths of the second to fourth digit (2D:4D ratio) was also calculated (see also Van Dongen, 2009). In addition, 19 landmarks were placed ( Fig. 1) and based on these landmarks, a procrustes analysis was performed in MorphoJ (available at: http://www.flywings.org.uk/ MorphoJ_page.htm; Klingenberg, 2011) to extract an overall measure of facial FA (faceFA2). In addition, facial masculinity was obtained as outlined in Little et al. (2008) and the EME angle was also calculated (Daniel & Pawlowski, 2007). Masculinity was also obtained from the procrustes analysis in MorphoJ by performing a canonical variate analysis for sexual dimorphism. This will allow to visualize the sexual dimorphism and to correlate the canonical variate with the measure obtained following Little et al (2008) and as outlined in Fig.1. An average measure of facial masculinity was obtained from the four individual measures after standardisation. All measurements were performed in ImageJ, freely available at http://rsb.info.nih.gov/ij/. First we tested if measures of masculinity differed between males and females using t-tests. The correlations among the masculinity measures (facial, EME angle, HGS, 2D:4D) were also graphically explored using a biplot from a principal component analysis. Next, correlations with FA, sexual behaviour and attractiveness were also investigated.

Literature search and meta-analysis
Studies investigating associations between HGS, attractiveness, FA, other forms of masculinity/femininity, sexual behaviour and dominance were obtained from Web of Science and PubMed. Six papers were found of which results are summarized in Table 1. Effect sizes (Pearson's correlations) from these studies as well as the results presented here were grouped in 5 different categories: masculinity measures (objective measurements on body or face); digit ratios; ratings (of masculinity, dominance, popularity); sexual behaviour (age of first contact, promiscuity) and attractiveness. Effect sizes in these categories and for males and females were presented in a funnel plot (i.e., in relation with sample size) to explore problems of publication bias. Effect sizes were then compared among the 5 categories and between males and females by a mixed model ANOVA with reference as random effect.

Measures of masculinity/femininity and sexual dimorphism
Each of the four measures of masculinity (based on the landmarks in Fig. 1) showed a statistically significant sexual dimorphism (Table 2). Therefore, an average measure was obtained after standardisation (further called facial masculinity or masc_face Lower face height / face height 0.59 0.57 t 97 =-3.14 0.002 Table 2. Tests of facial sexual dimorphism in the four individual measures (Fig. 1).
Facial shape also differed significantly between males and females based on the geometric morphometrics approach (p<0.0001). The shape differences are given in Figure 2. The canonical variate of the shape difference between males and females correlated strongly with facial masculinity as calculated above (see Table 2 for details). Thus, facial masculinity as measured by the relative proportions of different distances in the face (Table 2; Little et al., 2008) closely reflects the sexual dimorphism present in the landmarks used. We, therefore, used facial masculinity based on the relative proportions of the distances in Figure  1 for comparability with other studies. Across males and females:  Table 3. Correlation coefficients and statistical significance (*: p<0.05; **: p<0.01; ***: p<0.001, indicated in bold) of associations among fluctuating asymmetry (FA) values (hand and face), measures of masculinity (face, hand grip strength (HGS), eye-mouth-eye angle and digit ratio (2D:4D)), sexual behavior (number of partners and age of first sexual contact (AFC)) and attractiveness. Correlations are given across both sexes (top table) and for males and females separately (bottom table). Next to the facial masculinity studied here (which was significantly dimorphic: t 97 =5.69, p<0.0001), only one other measure of masculinity/femininity also showed a significant sexual dimorphism in our sample. Males showed significantly higher HGS (t 97 =10.1, p<0.0001), but no differences were observed for 2D:4D (t 94 =1.54, p=0.12) and the EME angle (t 97 =0.92, p=0.36) (see also Fig. 4). Across males and females, only HGS and facial masculinity showed a significant positive correlation (Table 3), a pattern that appeared consistent among both sexes (Fig.5), albeit not significantly so within sexes (Table 3). In woman, the EME angle and facial masculinity were negatively correlated, yet, unexpectedly, slightly positively in males (Table 3, Fig. 5). A principal component analysis of the  (Table 3).

www.intechopen.com
Hand Grip Strength in Relation to Morphological Measures of Masculinity, Fluctuating Asymmetry and Sexual Behaviour in Males And Females 301 masculinity traits confirmed the lack of strong correlations. The first principal component explained 37% of the total variation and was determined by HGS and facial masculinity. The second and third each explained about 25% of the total variation and each reflected one other variable, 2D:4D and EME angle respectively (Fig.6). Thus, in order to capture a large amount of the variation, three components were required, one of which only combined variation across two variables.

Asymmetry measurements
The degree of measurement error (ME) of FA (i.e., the percentage of variance due to ME relative to the total variance (FA+ME)) due to scanning and placing landmarks, were the following: D2: 16%; D3: 20%; D4: 18%; D5: 13%. None of the hand measurements showed significant directional asymmetry (all p > 0.05). After standardization, asymmetries of handtraits were averaged into a single measure of asymmetry per individual (FAhand). Both handedness and asymmetry in power between right and left hand were significantly correlated with the signed asymmetry of the hand (handedness: r = 0.33, p = 0.001; power: r = 0.30, p = 0.001) (see also Van Dongen et al., 2009). For the three facial characteristics measurements were less accurate (EW: 57%; EN: 14%; CM: 73%). Facial FA showed significant directional asymmetry (all p > 0.05), two facial traits (EW: t 99 = 3.2, p = 0.002 and PN: t 99 = 1.98, p = 0.05) showed larger values on the right side, on average. Because the three measurements of facial FA did not show high accuracy (though did show associations with sexual behavior, see Van Dongen et al., 2009 and below) we decided to take additional measurement on the face in the form of 19 landmarks. Procrustes ANOVA showed significant directional asymmetry (F 31,1798 =2.42, p<0.0001) and significant FA (F 1798,1426 =2.28, p<0.0001).

Correlations among FA, masculinity and sexual behaviour
All correlations, across both sexes and for males and females separately, are provided in Table 3. It is important to realize that many tests are being performed and some of them are significant at the 5% level just by chance. It is, therefore, only relevant to interpret correlations significant at the 5%/135=0.04% level (after Bonferonni correction). The only correlations which are significant at this level are situated between FA and sexual behaviour, between the number of partners and age of first sexual contact, between the two measures of facial FA, and between facial masculinity and HGS (Table 3). One correlation that is worth mentioning (albeit not significant after Bonferonni correction) is the negative association between EME angle and facial FA in woman (Table 3) indicating a wider (more feminine) EME angle to be associated with higher facial FA (Fig.7).

Meta-analysis
A funnel graph of all available effect sizes is provided in figure 8. There does not appear to be a problem of publication bias (correlation between sample size and effect size = -0.11, d.f.=81, p=0.29). 16 out of 83 estimates were statistically significant (20%) and 64 out of 83 estimates were in the expected direction (i.e., a positive effect size) (77%), a proportion that is significantly higher than 50% (p<0.0001). The average weighted effect size across all estimates equalled 0.19 (0.05), which was significantly different from zero (t 7 =4.00, p=0.007). Thus, on average there appears to be a robust correlation. However, average effect sizes were only half as high in females (difference=-0.10 (0.03), F 1,74 =11.3, p=0.001), and differed significantly among the broad categories of masculinity and sexual behaviour (F 1,66 =3.06, p=0.015). Although most two-by-two comparisons were not statistically significant, averaged across males and females, the highest effect sizes that were significant at the 0.01 level were found for objective measurements of bodily and facial masculinity (average effect size: 0.24 (0.05)) and ratings of dominance and attractiveness of opposite sex raters (average effect size: 0.22 (0.05)). Lower effect sizes, albeit still significant at the 0.05 level, were observed for 2D:4D (average effect size: 0.13 (0.05)) and measures of sexual behaviour (average effect size: 0.14 (0.05)). The remaining two were even somewhat lower, no longer statistically significant but still in the expected direction: attractiveness (average effect size: 0.12 (0.08)) and self rated dominance, aggression and popularity (average effect size: 0.09 (0.05)). These differences appeared comparable between males and females as there was no significant interaction (F 5,68 =2.00, p=0.09), but the power to detect an interaction was probably small. Therefore, we also present average effect sizes by sex (Table 4) Table 4. Average weighted effect sizes of the associations between hand grip strength and other measures of masculinity (masculinity: objective measurements; ratings: ratings of masculinity and dominance by opposite sex raters; self ratings: own evaluations of masculinity, dominance, popularity; digit ratios (2D:4D), attractiveness and sexual behavior) for males and females. Categories where males have a significantly higher effect size are indicated by a * (although the interaction was not statistically significant, see text for details).

Associations between masculinity, attractiveness, fluctuating asymmetry and sexual behaviour
This study, albeit small in terms of new data added to the existing literature, did not provide s t r o n g e v i d e n c e t h a t m e a s u r e s o f m a s c u l i n i t y w o u l d b e r e l a t e d t o s e x u a l b e h a v i o u r , attractiveness or fluctuating asymmetry. Clearly, sample sizes were relatively small, yet, it did allow to detect robust associations between FA and measures of sexual behaviour (see Van Dongen et al., 2009 for further discussion), but not attractiveness (this study). Thus, this suggests that sample sizes were sufficiently large for some aspects (i.e., associations with FA), and that asymmetry may be more closely related to sexual behaviour and promiscuity that masculinity. Nevertheless, many others have shown associations between masculinity and both attractiveness and sexual behaviour, such that this small study clearly cannot cast any doubt on the relevance of masculinity and hormone levels in human sexual behaviour and attractiveness. However, there is some doubt about the associations among different measures of masculinity and their association with sex-hormone levels (e.g., Koehler et al., 2004;Campbell et al., 2010). In this study, associations among the four objective measures were weak, with the exception of the association between facial masculinity and hand grip strength. HGS also showed a clear sexual dimorphism, as did facial masculinity. However, eye-mouth-eye angle and 2D:4D did not show correlations with facial masculinity or HGS and were not sexually dimorphic. Results for 2D:4D are discussed elsewhere (Van Dongen 2009).
For EME angle, one study of similar size as this one did show a sexual dimorphism and associations with attractiveness (Danel & Pawlowski, 2007). The results presented here thus question the generality of the usefulness of EME angle as a measure of masculinity and calls for further research. In spite of the fact that EME angle did not show a sexual dimorphism in this study and did not relate to masculinity (except perhaps weakly in woman), sexual behaviour or attractiveness, there was some suggestion that it correlated with facial FA. This certainly warrants further study since associations between FA and measures of masculinity are at best very weak and results vary among studies (Van Dongen submitted manuscript).

Handgrip strength as a measure of masculinity
Our results show that HGS relates to facial masculinity (but not 2D:4D) in both males and females. HGS has only recently been put forward as a useful measure of masculinity (Table 1), and we here present an overview of the current literature. There appears to be a highly significant and robust average weighted effect size of about 0.2, of correlations between HGS and different correlates of masculinity/femininity. There also appears to be some variation in the effect sizes. On the average, effect sizes were smaller in females and lowest for sexual behaviour and self rated dominance, aggression and popularity. Although there was no significant interaction between sex and type of masculinity measure, the p value was only 0.09, suggesting that the difference between may not have been similar for the different categories.
Although we should interpret these test with caution (and await further study), the differences in effect sizes between males and females were strongest for attractiveness, sexual behaviour and rated masculinity. In each of these, relatively strong average effect sizes were observed for males, and nearly zero for females (Table 4). Thus, HGS appears to be related to objectively measured masculinity in both males and females (Table 4 and data from this study), and to a lesser extent with 2D:4D (Table 4). For all other categories, no significant associations were found for females (Table 4). Although it may be to preliminary at this point to make any firm conclusions, our results and the combined analysis of the data from the literature suggests that HGS relates to morphological measures of masculinity alone in females, but also to attractiveness, rated and self-rated masculinity and dominance and sexual behaviour in males.

Conclusion
In this paper we study associations between objective morphological measures of masculinity/femininity and physical strength (handgrip strength) in relation to developmental instability (as measured by fluctuating asymmetry, FA), attractiveness and sexual behaviour. In spite of the relatively small sample sizes, we were able to detect associations between FA and sexual behaviour (further discussed in Van Dongen et al., 2009), yet not with our measures of masculinity. We next focussed on a relatively recently studied measure of masculinity/femininity, namely physical strength expressed as handgrip strength (HGS). We reviewed results from the recent literature and demonstrated a robust association between HGS and other measures of masculinity/femininity. In addition, we were able to detect some sources of variation. On the one hand, HGS related to morphological features of bodily masculinity (and to a lesser extent but still significantly so to 2D:4D ratios) equally strong in both males and females. However, associations between HGS and either attractiveness, (self-)ratings of dominance, masculinity and popularity and sexual behaviour were weaker or absent in females compared to males. Thus, based on the available literature we conclude that physical strength is determined by circulating hormones affecting morphologically dimorphic structures, yet affects behaviour and the physical expression of it in males only. Physical strength and masculinity is thus likely to play a role in male-male competition and as a signal of mate value in sexual selection.